2 Messages
•
84 Points
What's the right way to report an AI-powered profile posting slop reviews?
https://www.imdb.com/user/ur181362155/ is an account that posts AI slop reviews by the thousands.
I've reported several, yet every one of them was declined for "Unable to verify" reasons.
Here's an example: https://www.imdb.com/review/rw10117649/
Every single review is of that format; if you look at a handful of them, it's impossible not to see it. Everything is described with two adjectives. There are usually no solid details, just stuff an AI might pick up from being trained on some early (pre-2024) reddit threads about the show, not this particular episode. Even more, when it does come up with a solid detail (Louise finds a creepy doll in the basement) it is a hallucination that anybody who has watched the episode would pick up on very quickly.
This should be obvious from the user's account history. It has posted over seven hundred reviews per day since it was created.
Does IMDB have an official policy on AI slop? If garbage, low-quality, hallucinatory reviews are not actually allowed, how does it expect those to be surfaced if not by the reporting system?
gdorn
2 Messages
•
84 Points
1 month ago
Here's ChatGPT, after I fed it the show's title and an arbitrary rating of 7. Importantly, I did not feed it the description, and it hallucinated finding something in the basement, which, again, is not something that happens in the episode. Also check out the vacuous word choices that, sorry, no human uses this way.
It replies:
0
Fran
Employee
•
2.7K Messages
•
27.4K Points
1 month ago
Hello gdorn
Thank you for reporting this issue. I've forwarded this information to the appropriate team for further investigation (Ref Ticket #V1604209308). I'll reply once we receive further information. Thank you for your patience.
0